Handbook on Intelligent Healthcare Analytics. Группа авторов

Чтение книги онлайн.

Figure 2.2 Proposed system for predicting disaster using improved Bayesian hidden Markov frameworks (IBHMF).

2.4 Results and Discussion

      From the observations on each state moves, the probability with the power of the matrix is generated. Let us consider the steps as state-to-state transition.

Let A and B be the two states with K steps, which have many possibilities to initialize the time with constant value. When the iterations started from one to the i-th iteration, there will be the same number of vector probability ratio that are identified with three states such as natural disasters (Figure 2.4), dry, and drought. Figure 2.3 shows the variations of number of disasters.

Table 2.2 defines the dataset year-wise and its ratio when it is trained according to the NN. In addition, testing data with rating as well reported on a year basis are monitored from the parameters that match according to the testing and training accuracy.

      Figure 2.4 shows the entities that are identified from the united states of America on various years that have the impact according to the climate changes. Prediction using this dataset based on the developed feature analysis can be performed. Each and every disaster along with the disasters reported year-wise with the count is also reported.

Natural disaster according to the disaster effects and probability of occurrence using stochastic matrix is described in the analysis. Figure 2.5 illustrates the understanding of probability ratio. It clearly indicates the year-wise analysis of the disaster along with the economic damage that is caused by these damages. Figure 2.6 shows the boxplot view of natural disasters on various entities.

Figure 2.3 Total number of disaster analysis using improved Bayesian Markov chain model.

2.5 Conclusion

As observed from the Markov chain model that is developed, a framework for identifying random effects can create a disaster. Kaggle datasets are used to find the state transition matrix and occurrence that are predicting the natural changes in climate. Computation on the basis of frequency and the latent state changes are monitored based on the probability. By considering the likelihood from the part of observation that are identified from the matrix according to rows and column for assumption. Using the sequence numbering and entries on hidden state, the current sequence follows the Markov model and fixes the likelihood. Using the proposed framework, the tasks such as important performance based on weather forecast are achieved. The proposed algorithm IBHMF that produced a better performance and also independent variables that are factors of time series shows the exact transition matrix analysis to predict the forecast based on climatic change.

Table 2.2 Sample dataset for predicting weather forecasting.

	Year	Total economic damage from natural disasters (US$)
count	561.000000	5.610000e+02
mean	1977.217469	1.146966e+10
std	30.399233	3.199525e+10
min	1900.000000	0.000000e+00
25%	1959.000000	6.50000e+07
50%	1984.000000	8.400000e+07
75%	2001.000000	5.444777e+09
max	2018.000000	3.640932e+11

Figure 2.4 Changes from various impacts from natural disaster.

Figure 2.5 Economic damage changes a prediction analysis.

Figure 2.6 Boxplot view of natural disaster on various entity.

References

      1. Baboo, S.S., Baboo, S.S., Shereef, I.K., An Efficient Weather Forecasting System Using Artificial Neural Network. Int. J. Environ. Sci. Dev., 1, 321–326, 2010, https://doi.org/10.7763/ijesd.2010.v1.63.

      2. Cognitive Tasks and Learning, n.d. SpringerReference, https://doi.org/10.1007/springerreference_226188.

      3. Li, D. and Yu, D., Deep Learning: Methods and Applications. Found. Trends Signal Process., 7, 3–4, 197–387, 2014.

      4. Grinsted, A., Ditlevsen, P., Christensen, J.H., Normalized US Hurricane Damage Estimates Using Area of Total Destruction. Proceedings of the National Academy of Sciences of the United States of America, vol. 116, pp. 23942–46, 20191900-2018.

      5. Anand, J. et al., Efficient Security for Desktop СКАЧАТЬ